Journal of General Virology (1998), 79, 673–681.

Printed in Great Britain

...................................................................................................................................................................................................................................................................................

Nucleotides in the panhandle structure of the influenza B virus

virion RNA are involved in the specificity between influenza A
and B viruses
Yeon-Sook Lee and Baik L. Seong
Institute of Biological Sciences, Hanhyo Institutes of Technology, 461-6 Chunmin-dong, Yoosung-gu, Taejun, 305-309, South Korea

Influenza A and B viruses share common sequences duplex region were utilized by influenza B virus,
and potentially similar panhandle structures in the whereas these mutations were detrimental for
terminal noncoding regions of virion RNA (vRNA). influenza A virus. Different activity profiles were
Interesting differences exist, however, in the num- observed in the length preference of the RNA
ber of conserved nucleotides at the 5« and 3« ends of duplexes : maximum template activity was observed
the vRNAs, in base pairs constituting the panhandle with 11 base pairs for influenza B virus, and 8 base
duplex, and the length of uridine stretch (U stretch) pairs for influenza A virus. When the mutants with
juxtaposed to the RNA duplex. To analyse the various lengths of U stretch were tested, highest
contribution of these signals to the specificity CAT activities were observed with 5 to 7 uridine
between the two viruses, a transient ribonucleo- residues in influenza A virus, whereas in influenza B
protein transfection method was used for the ex- virus the activity was drastically decreased with 7
pression of the chloramphenicol acetyltransferase uridine residues. We suggest that the specific inter-
(CAT) reporter gene flanked by the noncoding action of influenza virus RNA polymerase with these
nucleotides derived from influenza B vRNA. While noncoding cis-acting signals in transcription of the
the base pairing in the RNA duplex was primarily RNA genome, along with unique coding strategies
important for template activity, mismatch mutations adopted by influenza B virus, has contributed to the
G11¬G12« and C12¬A13« in the terminal RNA divergence of these two closely related viruses.

Introduction Moreover, differences between the evolutionary pattern of A

Influenza A, B and C viruses are a group of enveloped and B type viruses are also documented (Smith & Palese, 1989 ;
negative-sense RNA viruses characterized by their segmented Yamashita et al., 1988).
genome. Among these, A and B types are important human Like influenza A viruses, influenza B virus contains eight
pathogens (Murphy & Webster, 1990). Although serious segments of single-stranded RNA of negative polarity, and
infections by influenza B virus are less well documented than each RNA segment is flanked by conserved sequence elements
are influenza A virus infections, influenza B virus remains the (Stoeckle et al., 1987). The noncoding sequences at the 3« and
major cause of illness associated with influenza infections the 5« ends of virion RNA (vRNA) are partially complementary
among school children (Hall et al., 1979 ; Kim et al., 1979 ; to each other and can form an extended panhandle structure.
Nolan et al., 1980). Influenza B virus exhibits unique diversity Despite apparent similarity in genomic structure, there are
in coding strategies (Horvath et al., 1990 ; Lamb & Choppin, some interesting differences between influenza A and influenza
1983 ; Lamb & Horvath, 1991 ; Shaw et al., 1983, 1992), B virus genomic RNA. First, the base-paired duplex region of
biological properties and host ranges (Kingsbury, 1990). influenza B virus is usually longer than influenza A virus (8–10
base pairs in influenza B virus and 4–8 base pairs in influenza
A virus). Second, the uridine (U) stretch, a potential poly-
Author for correspondence : Baik L. Seong. adenylation signal, usually consists of 5–7 uridine residues in
Fax ­82 42 866 9128. e-mail blseong!donald.hanhyo.co.kr influenza A virus, whereas in influenza B virus the number of
uridine residues is 5–6, and no 7 uridine stretches are found.
Part of this work was presented in the conference ‘ Options for the Third, the conserved nucleotides at the 3« and 5« ends of vRNA
control of influenza III ’ in May 1996, in Cairns, Australia. are 9 and 10 nucleotides, respectively, in influenza B virus, and

0001-5207 # 1998 SGM GHD

Y.-S. Lee and B. L. Seong
12 and 13 nucleotides, respectively, in influenza A viruses.
Besides, the position containing natural variations within the
conserved region is also different between the two viruses.
Due to recent breakthroughs in reverse genetics tech-
nology, much progress has been made in the understanding of
RNA signals involved in transcription and replication of the
influenza A virus RNA genome (Bergmann & Muster, 1996 ;
Flick et al., 1996 ; Fodor et al., 1995, 1996 ; Hagen et al., 1994 ;
Kim et al., 1997 ; Li & Palese, 1994 ; Luo et al., 1991 ; Neumann
& Hobom, 1995 ; Piccone et al., 1993 ; Yamanaka et al., 1991).
Various methods and their modifications have been described
for a reverse genetics system for influenza A virus. They
include transfection of in vitro-reconstituted ribonucleoprotein
(RNP) complex into helper influenza virus-infected cells (Enami
& Palese, 1991 ; Luytjes et al., 1989 ; Martı! n et al., 1992 ; Seong
& Brownlee, 1992 a) and expression of virus-like RNAs and
viral proteins in cultured cells using plasmids or recombinant
viruses (Mena et al., 1994 ; Neumann et al., 1994 ; Zhang & Air,
1994). Using influenza proteins isolated from purified virions,
promoter analysis in vitro has sometimes been complicated due
to the presence of influenza-derived short RNA fragments in
the polymerase preparations (Seong & Brownlee, 1992 b ;
Fodor et al., 1995 ; Hagen et al., 1994 ; Tiley et al., 1994).
However, the same enzyme preparations have been suc-
cessfully used for transient transfection of RNP for analysis of
promoters}cis-acting signals involved in transcription of
influenza A RNA in cultured cells (Kim et al., 1997 ; Li & Palese,
1994 ; Luo et al., 1991 ; Piccone et al., 1993 ; Yamanaka et al.,
1991). Molecular aspects of influenza B virus RNA replication
are still not well understood. Moreover, though accumulated
gene sequence data suggest a common structure of vRNA
shared by both viruses, very little is known about the cis-acting
elements which determine the specificity between the two
viruses. In our previous work in vitro, influenza B virus
polymerase was shown to have a broader allowance for
sequence variations in the panhandle region than influenza A
virus polymerase (Lee & Seong, 1996). Extensive analysis, in
vivo and in vitro, of the ‘ RNA-fork ’ model of transcription of
influenza A vRNA showed that base pairing in duplex domain
II (see Fig. 1) plays an important role in RNA template activity
(Fodor et al., 1995 ; Kim et al., 1997). Since sequence divergence
between influenza A and B viruses is more apparent in domain
II rather than in domain I, we hoped to find signals influencing
the specificity of these two viruses by detailed analysis of this
part of the RNA duplex. To investigate the specificity and
divergence between these two related viruses in vivo, we used
the transient RNP transfection method in influenza virus-
infected cells with the CAT gene mimicking an influenza B
virus reading frame.
Methods
+ Viruses and cells. Influenza strain B}Lee}40 was grown in 11-day-
old embryonated eggs at 37 °C for 48 h and used as helper virus in a
transfection assay. B}HK}HG (a gift from Dr Laver, John Curtin Medical
GHE
School, Canberra, Australia) and B}Panama viruses (a gift from Evans
Medical Ltd, Liverpool, UK) were used in core preparation. A batch of
A}X-31 was purchased from Evans Medical Ltd. MDBK cells were grown
in DMEM enriched with 10 % foetal calf serum.
+ Preparation of RNP cores. Preparation of RNP from the A}X-31
virus was as previously described (Seong & Brownlee, 1992 a). RNP
preparation from B}HK}HG reassortant virus was according to Almond
et al. (1979) with slight modification. Briefly, concentrated egg-grown
virus was dissolved in disruption buffer containing 10 mM Tris–HCl
(pH 7±4), 100 mM NaCl, 100 mM KCl, 50 mM MgCl , 4 mM DTT, 1 %
#
NP40 and 0±5 % lysolecithin. After incubation for 30 min at 30 °C, the
lysate was fractionated on a glycerol step gradient of 70 %, 50 % and 30 %
in 10 mM Tris–HCl (pH 7±4), 100 mM NaCl, 100 mM KCl, 50 mM
MgCl and 4 mM DTT. Fractions enriched with nucleoprotein and
#
polymerase proteins were identified by SDS–PAGE and Coomassie blue
staining. RNP fractions were pooled and treated with micrococcal
nuclease (Seong & Brownlee, 1992 a) and subsequently used for
reconstitution of RNP.
+ Construction of plasmids. pBCAT-NS carries the CAT gene
flanked by the noncoding region of the nonstructural (NS) gene of the
B}Lee}40 virus. This plasmid was constructed by PCR using pIVACAT1
(Luytjes et al., 1989) as template with two primers : primer 1, 5«
AGCGGATCCGGTTAATACGACTCACTATAAGTAGTAACAA-
GAGAGGATTTTTATTTTACATTTACGCCCCGCCC 3« ; primer 2, 5«
AGCGGATCCGACGCCCTGCAGCAGAAGCAGAGGATTTATT-
TAGTCACTGGCAAACGGAAAGATGGACAAAAATCACTGGG
3«. The noncoding sequence of the NS gene is underlined. Mutagenesis
of the influenza B virus noncoding sequences was achieved by PCR using
pBCAT-NS as template and derivatives of the above primers carrying the
desired mutations. The mutations in all CAT constructs were verified by
dideoxy sequencing of plasmids.
+ RNP transfection to MDBK cells. The BCAT-NS RNA was
complexed with the influenza proteins derived from influenza B virus and
transfected into the B virus-infected MDBK cells. For comparison, the
same BCAT-NS RNA was reconstituted with influenza A virus-derived
proteins and transfected into the A virus-infected cells. RNP complexes
were made according to Enami & Palese (1991) using the micrococcal
nuclease-digested influenza virus cores as a source of nucleoprotein and
polymerase proteins (Seong & Brownlee, 1992 a). Briefly, 0±5 µg template
DNA, linearized with HgaI and filled with Klenow polymerase, was
transcribed in 1¬ transcription buffer (NEB), 2 mM each rNTP, 40 U
RNasin and 100 unit T7 RNA polymerase in the presence of 10 µl
micrococcal nuclease-digested core (0±5–1 µg nucleoprotein equivalent)
in 50 µl total reaction volume. Seventy to eighty percent confluent
monolayers of MDBK cells in 35 mm dishes were first infected with
helper virus (100–200 HAU per dish ; 10$–10% p.f.u. per HAU) for 1 h
prior to transfection with RNP. The DEAE–dextran method was used for
transfection. As helper virus, influenza A}X-31 or B}Lee}40 was used.
After transfection, 2 ml serum-free DMEM containing 10–20 µg}ml
trypsin was added.
+ CAT assay and quantification. After 12–20 h of incubation post-
transfection, cells were harvested and crude cell extracts were used for the
CAT assay (Gorman et al., 1982). Acetylated products were separated on
a silica thin-layer chromatography plate and exposed to X-ray film. The
expression levels were estimated by density scanning of the spots on the
X-ray film with LKB Ultroscan XL enhanced laser densitometer.

Results
Effects of base pair mutations in the panhandle
structure
In both influenza A and B virus RNA genomes, partially
complementary sequences at their termini allow the vRNA to
form a panhandle structure. However, the conserved sequences
in influenza B virus are confined within 9 and 10 nucleotides at
the 3« and 5« ends, respectively, whereas in influenza A virus,
the conserved sequences are 12 and 13 nucleotides. To test the
potential role of the difference in panhandle nucleotides, the
nucleotides and base pairs were systematically changed and
their effects on CAT activity were analysed using influenza A
or B virus as helper.
3«-11/5«-12 position. Nine different mutants either with base
mismatch or with base pair substitutions at the 11 : 12« position
were tested (Fig. 1). When influenza B virus was used as helper
and the source of polymerase proteins in RNP reconstitution,
most of the mismatch mutants were inactive (C11 ! A11 and
U11, lanes 1 and 7 ; G12« ! U12«, C12«, A12«, lanes 3, 6 and
9, respectively). However, strong CAT activity was observed
with the C11 ! G11 mismatch mutant (lane 4, 32 % of wild-
type). Reconstitution of the same BCAT-NS G11 mutant
template with influenza A proteins and transfection into
influenza A virus-infected cells did not produce any CAT
activity (data not shown). The detrimental effects of mismatch
mutations were compensated, at a low level, by restoring the
base pair [C11 : G12« ! A11 : U12« (lane 2), G11 : C12« (lane 5),
and U11 : A12« (lane 8) ; 5 %, 21 % and 10 % of wild-type,
respectively]. However, the CAT activity of the G11 : C12«
base-paired double mutant was still lower than that of the G11
mismatch mutant (compare lanes 4 and 5). High activity
towards the G11 mismatch mutant probably means that, in
influenza B virus, the base pairing at the 11 : 12« position is not
crucial for transcriptase activity. The sequence data of
B}HK}73 showed that the natural variation at position 11
sometimes did not involve base pairing with the corresponding
12« nucleotide (Desselberger et al., 1980). It should also be
remembered that the G11 mismatch was well tolerated by
influenza B virus polymerase in previous analysis in vitro (Lee
& Seong, 1996).
3«-12/5«-13 position. Fig. 2 (A) shows the effects of mutations
in the 12 : 13« base pair on CAT activity. Most single base
mismatch mutations, either at the 3«-12 or at the 5«-13 position,
were detrimental for CAT activity (U12 ! A12 and G12, lanes
1 and 4 ; A13« ! U13«, C13« and G13«, lanes 3, 6 and 9,
respectively). Exceptionally, the U12 ! C12 mutant with its
C12¬A13« mismatch exhibited strong CAT activity (lane 7,
41 % of wild-type). All base-paired double mutants were active
in CAT expression (U12 : A13« ! A12 : U13«, G12 : C13« and
C12 : G13« in lanes 2, 5 and 8 ; 88 %, 103 % and 29 % of wild-
type, respectively). The level of CAT activity by base pair
restoration was higher than the level observed with the
Specificity between influenza A and B viruses
Fig. 1. Effect of mutations at the 3«-11/5«-12 position on CAT expression
by influenza B virus. The panhandle can be divided into domain I and II
interrupted by a loop. The C11 and the G12« nucleotides (shown as bold
letters) were replaced by three other nucleotides with or without base
pairing. The activities of the mutants are presented as percentages of the
BCAT-NS wild-type activity. Notable differences are observed with the
G11¬G12« mismatch mutant (lane 4) between A and B viruses.
–, Activity below detection level.
adjacent 11 : 12« pair (Fig. 1). When the same BCAT-NS
mutants were rescued with influenza A virus, base pairing
ability was again crucial for CAT expression. Moreover, the
C12 mutant with the C12¬A13« mismatch was slightly active
in the CAT assay (6 % of wild-type ; data not shown). The
results are consistent with similar analyses with influenza A
virus using the IVACAT1 RNA containing the noncoding
sequence derived from A virus (Kim et al., 1997). The results
suggest that at the 12 : 13« position, with the potential
exception of the C12¬A13« pair, the base pairing ability,
rather than the nature of the nucleotide, is the important
determinant for template activity.
During the course of our studies on template specificity of
influenza A and B viruses, we observed that the IVACAT1 was
nearly inactive with influenza B virus (Fig. 2 B, lane 2). Since the
U12 : A13« base pair is present in the majority of influenza B
vRNA segments, we replaced the C12 : G13« base pair
conserved in the IVACAT1 RNA template with the U12 : A13«
base pair. This mutant IVACAT1 exhibited low but clearly
detectable activity using influenza B virus as helper (9 % of
wild-type BCAT-NS ; Fig. 2 B, lane 3). The results showed that
influenza B virus polymerase had a preference for the
U12 : A13« pair over the C12 : G13« pair. Previous analyses
with influenza A virus with the homologous IVACAT1
template showed that the C12 : G13« base pair was threefold
more active than the G12 : C13« base pair (Kim et al., 1997). The
opposite preference of A and B viruses suggests that the nature
of nucleotide at the 12 : 13« position may be involved in the
specificity of the influenza B virus polymerase for the A and B
virus RNA templates.
13 : 14« and 14 : 15« positions. None of the mismatch mutants in
the 13 : 14« base pair, either by the 3« or the 5« mutations, were
GHF
Y.-S. Lee and B. L. Seong
Fig. 2. (A) Effect of mutations at the 3«-12/5«-13 position on CAT
expression. The 12 : 13« base pair to be mutated is shown in bold. The
activities of mutant templates are presented as percentages of the BCAT-
NS wild-type activity. The different activity between A and B viruses is
found with the C12¬A13« mismatch mutant which is active only with
influenza B virus (lane 7 ; see text). (B) Replacement of the C12 : G13«
base pair of the IVACAT gene (lane 2 ; sequence shown above) (Luytjes et
al., 1989) with the U12 : A13« pair (lane 3). IVACAT RNAs, which carry
influenza A-derived noncoding sequences, were complexed with influenza
B virus proteins and transfected into influenza B virus-infected cells.
active as templates (Fig. 3 A, lanes 1, 3, 4, 6, 7 and 9). In all
cases, however, the CAT activity was rescued by base pair
formation (C13 : G14« ! A13 : U14« (lane 2), G13 : C14« (lane
5), U13 : A14« (lane 8) ; 10 %, 87 % and 43 % of wild-type,
respectively). Limited numbers of mutants were also generated
at the 14 : 15« base pair. As shown in Fig. 3 (B), base pairing was
also critically important for CAT activity (C14 : G15« !
G14 : C15« (lane 2) ; 109 % of wild-type), whereas the single
base mismatch mutations were detrimental (C14 ! G14 in
lane 1 and G15 ! C15« in lane 3). All mutant BCAT-NS
templates at positions 13 : 14« and 14 : 15« were also recon-
stituted with influenza A proteins and transfected into influenza
A virus-infected cells. The activity profile was similar to Fig.
GHG
Fig. 3. Effect of mutations at the (A) 3«-13/5«-14 and (B) 3«-14/5«-15
positions on CAT expression. The mutated base pairs are shown in bold. In
both positions, stable base pairs were essential for template activity (A,
lanes 2, 5 and 8 ; B, lane 2).
3 (A, B) (data not shown). Therefore, no crucial differences in
template specificity were observed between influenza A and B
viruses.
From analyses of mutants of each base pair, 11 : 12«, 12 : 13«,
13 : 14« and 14 : 15« of the BCAT-NS, our results show that (1)
a stable base pair, rather than the nature of nucleotide, is
important in the central region of the duplex (e.g. 12 : 13«,
13 : 14« and 14 : 15«), (2) the nature of the nucleotide as well as
the base pair itself becomes important in the distal part of the
duplex (e.g. the 11 : 12« base pair), (3) some mismatch mutations
at the 11 : 12« and 12 : 13« positions are tolerated with B virus
but not with A virus, and (4) the 12 : 13« base pair may play a
 Specificity between influenza A and B viruses

role in the difference in specificity of the A and B virus RNA

templates.

Effect of panhandle length

Generally, the length of the base-paired RNA duplex in
influenza B virus is 8–10, slightly longer than that of influenza
A virus, which ranges from 4 to 8. To investigate the potential
effect of its length on template efficiency, we generated various
mutants with different length RNA duplex regions, either by
direct deletion of base pairs within the duplex or by introducing
mutation(s) at the 3« side of the distal end of the panhandle,
effectively increasing the number of base pairs.
Direct deletion of the U15 : A16« base pair resulted in
dramatic decrease in CAT activity (about 8 % of wild-type).
Deletions of 2 base pairs (14 : 15« and 15 : 16«) or 3 base pairs
(13 : 14«, 14 : 15« and 15 : 16«) further enhanced the deleterious
effect, resulting in no detectable activity (data not shown). The
sharp decrease in CAT activity could be due either to the
shortening of the duplex length or to the change in distance
between the 5« termini and the U stretch, which has been
shown to be an important factor for polyadenylation in
influenza A virus (Li & Palese, 1994). To ascertain this, we
compared mutants carrying various lengths of the duplex
ranging from 4 to 14 base pairs which were generated by
cumulative mutations from the distal end of the RNA duplex
of the BCAT-NS RNA (Fig. 4 A). As shown in Fig. 4 (B), the
highest activities were observed within the range of 8–13 base
pairs (85–150 % of wild-type) with influenza B virus, with the
peak activity with 11 base pairs. When influenza A virus was
used as helper, however, the activity profile was different. The
activity was almost saturated within 6–9 base pairs (83–128 %
of wild-type) with the peak activity with 8 base pairs. It should
be noted that, with influenza A virus as helper, good CAT
activities were observed with the 8 and 7 base pair mutants,
which are 1 and 2 base pairs shorter than the 9 base-paired
wild-type, BCAT-NS, respectively.
This suggests that the detrimental effect caused by direct
deletion of 1 or 2 base pairs within the duplex is due to the
change in distance between the 5« termini and the U stretch
rather than to the shortening of the duplex length, which
emphasizes the requirement of the minimal distance between
the two (Li & Palese, 1994). Our new finding on the apparent
preference of influenza B virus for a longer RNA duplex in the
panhandle is consistent with the observation that the base-
paired duplex region of influenza B virus RNA is usually longer
Fig. 4. Effects of panhandle length on CAT expression by influenza A and
than influenza A virus RNA. B viruses. (A) The length of base pairs of the BCAT-NS RNA was changed
by cumulative mutations within the C14–U23 nucleotides at the 3« end
Effect of the length of U stretch (shown in bold), either disrupting or introducing base pairs. (B) The
average values of more than three independent experiments are presented
The length of uridine sequence adjacent to the RNA duplex as percentages of the 9 base-paired BCAT-NS wild-type.
is usually 5–7 nucleotides in A virus, whereas only 5 or 6
residues are present in influenza B virus (Akoto et al., 1987 ;
DeBorde et al., 1988 ; Desselberger et al., 1980 ; Kemdirim et al., uridine residues were constructed (Fig. 5 A). When infected
1986 ; Robertson, 1979 ; Stoeckle et al., 1987). To test its with influenza A virus, the highest CAT activities were
potential effect, a series of mutant BCAT-NS RNAs with 3–8 observed within the range of 5–7 uridine residues (Fig. 5 B).

GHH
Y.-S. Lee and B. L. Seong
Fig. 5. Effect of the U stretch on CAT expression by influenza A and B
viruses. (A) The U stretch of the 5« noncoding region of the BCAT-NS RNA
was changed by addition (shown in bold) or deletion (shown as ∆) of
uridine nucleotide(s). (B) While influenza A virus exhibited optimal activity
with 5, 6 or 7 uridine residues as previously reported (Li & Palese,
1994), influenza B virus did not show comparable activity with the 7
uridine mutant. The averages of three independent experiments are shown
as relative to the BCAT-NS wild-type carrying 5 uridines.
The 7 U mutant was almost 90 % as active as the 6 U mutant.
The results are consistent with similar work done with the
IVACAT1 template with influenza A virus (Li & Palese, 1994).
With influenza B virus as helper, however, the prime activities
were observed only with 5 and 6 uridine residues (100 and
130 %, respectively) and the activity critically declined with 7
uridine residues (8 % of wild-type). Therefore, influenza B virus
showed a narrower requirement than A virus for the number of
uridine residues. The 3 U mutant, generated by deletion of 2
uridine residues from the wild-type template, concomitantly
extended the number of base pairs at the RNA duplex (13 base
pairs). The mutant template with the same number of base
pairs, without deletion of the U stretch, exhibited as a strong
CAT activity as wild-type (Fig. 4 A, B). Therefore, the low
GHI
Fig. 6. Effect of point mutations at positions 4 and 10 of the 3« end. All
possible base substitution mutations were introduced at position 4 or 10
at the 3« end of the BCAT-NS RNA. The mutant templates were mixed with
influenza B virus core proteins and transfected into influenza B virus-
infected cells. A U or C transition is observed at position 4 in influenza A
virus, and at position 10 in influenza B virus. At both positions, significant
activities were observed only with U or C nucleotides (lanes 2 and 4) and
not with G or A nucleotides (lanes 3, 4, 6 and 7).
activity with the 3 U mutant could be directly ascribed to the
shortening of the poly U stretch.
Other mutations
In influenza A virus natural variation, U or C is observed at
position 4 at the 3« end, whereas, in influenza B virus, a similar
U or C variation is observed at position 10. To test potential
nucleotide specific effects, all three base substitution mutants
were generated at both position 4 and 10 (Fig. 6). Only U4 !
C4 and U10 ! C10 mutants exhibited significant, but lower
CAT activity (42 % and 10 % of wild-type, respectively). Base-
paired double mutants (C10 : G11« or G10 : C11«) did not rescue
the CAT activity (data not shown). In influenza B virus RNA
segments 1, 2 and 3, which encode the polymerase proteins,
the C10 nucleotide is present within the C10¬A11« mismatch.
The apparent activity with the C10 transition mutant without
a base pair (Fig. 6) and the lack of activity of base-paired double
mutants suggests that while the 10 : 11« base pair is not
essential for template activity, the C10 nucleotide may be
involved in down-regulation of polymerase proteins in
influenza B virus.
Discussion
The present study on the role of the nucleotides in the
panhandle region of the BCAT-NS gene suggests that,
primarily, base pairing within the RNA duplex is an important
factor for influenza B virus transcriptase activity. The com-
parative transient transfection analysis using A or B viruses,
however, revealed interesting differences between these two

related viruses. We observed that the length of RNA duplex,
the nature of base pairs in the RNA duplex and the length of
the U stretch all contributed towards differences in specificity
between the two viruses.
The greater tolerance of influenza B virus towards mismatch
mutations within the RNA duplex is consistent with earlier
observations in vitro (Lee & Seong, 1996). Moreover, mis-
matches are actually observed at the 11 : 12« and 12 : 13«
positions in some natural influenza B isolates (Desselberger et
al., 1980). The interpretation, however, is based on the
assumption that these nucleotides are part of the elongated
double-stranded domain (see Fig. 1). Whether this structure is
the only relevant conformation throughout the infectious cycle
remains to be determined. Since transfected CAT RNA is
expected to undergo multiple rounds of replication, the
proposed mismatches in the vRNA template would result in
mismatches in the cRNA at the corresponding positions and
affect vRNA synthesis as well. Therefore, the observed
differences between A and B viruses would be the composite
of effects on transcription, replication and potentially on the
vRNA}cRNA pool size ratios in infected cells.
The preference of B virus for the U12 : A13« pair over the
C12 : G13« pair was the opposite to that for influenza A virus
(Kim et al., 1997). Since the U12 : A13« pair is observed in
natural influenza B isolates and not in influenza A isolates, the
presence of this particular base pair in the influenza B virus
RNA template may have a detrimental effect on the interaction
of the influenza A virus polymerase with influenza B virus
RNA template. Moreover, heterogeneity of sequences in this
position observed among different influenza B virus RNA
segments may contribute towards the segment-specific regu-
lation of influenza B virus gene expression. The role of
discriminator nucleotides in the specificity among pools of
similar RNAs and RNA binding proteins has been well
documented in tRNA–synthetase interactions (Saks et al.,
1994). The proposed role of this base pair was further
supported by the inactivity of IVACAT1 RNA template with
B virus and partial rescue of CAT activity by replacement of
the C12 : G13« pair in the IVACAT1 template by the U12 : A13«
base pair. However, this base pair alone was not sufficient ;
other structural feature such as the number of base pairs in the
RNA duplex and the length of the U stretch (see below) were
also involved in the discrimination between the A and B virus
RNA templates.
The sequence variation observed naturally within the first
12 conserved nucleotides at the 3« end of the influenza A virus
RNA is the U or C variation at position 4. The C4 nucleotide
has been thought to be involved in the segment-specific down-
regulation of the polymerase genes, since the nucleotide has
been invariably documented within the polymerase gene
segments (Desselberger et al., 1980 ; Robertson, 1979). In
contrast, sequence conservation at the 3« end of the influenza
B virus RNAs is confined within the first 9 nucleotides and the
U or C variation is observed at position 10 (Stoeckle et al.,
Specificity between influenza A and B viruses
1987). According to our results, CAT activities were observed
with the U and C nucleotides but not with G or A. The three
polymerase genes in influenza B virus carry the C10 nucleotide,
and therefore, the lower activity observed with the C10
mutant template than that with the U10 template may
contribute, in part, towards the segment-specific down-
regulation of polymerase genes in some influenza B viruses.
Considering the role of some other nucleotides in the molecular
ends of the vRNA, as recently advanced in the ‘ cork-screw
model ’ of transcription of influenza A virus (Neumann &
Hobom, 1995 ; Flick et al., 1996), more extensive analysis of
these parts of the promoter element is required for better
understanding of transcriptional control of influenza B virus as
well.
In our analysis of base pair length mutations, A18 : U19«
and A17 : U18« base pairs did not contribute significantly
towards template activity. Therefore, four consecutive base
pairs, 13 : 14«, 14 : 15«, 15 : 16« and 16 : 17«, were essential,
although not sufficient, for transcriptase recognition of
influenza B virus RNA template. Influenza A and B viruses
exhibited slightly different profiles in their preference for the
length of the RNA duplex : in B virus, maximal template
activity was observed with 11 base pairs, whereas, in A virus,
the peak activity was with 8 base pairs. The preference for a
longer RNA duplex in B virus was consistent with accumulated
sequence data that natural influenza B virus isolates usually
carry a larger number of base pairs than influenza A virus
(Stoeckle et al., 1987). We observed that BCAT-NS mutants
with reduced numbers of base pairs (8–7 base pairs), but
keeping the same distance between the 5« end and the U
stretch, still showed strong CAT activities. However, mutant
templates with the same reduced number of base pairs which
concomitantly shortened the distance between the 5« end and
the U stretch exhibited greatly reduced CAT activity (! 10 %
of wild-type). This suggests that the distance between the 5«
end of the RNA template and the U stretch of the B virus RNA
template should be at least 16 nucleotides, as in A virus. The
effect of distance on template activity could, perhaps, be best
explained by the polyadenylation model recently proposed in
influenza A virus (Tiley et al., 1994 ; Fodor et al., 1996). The
steric hindrance proposed in this model would depend mainly
on the distance between the 5« end and the position of U
stretch of the RNA template.
Differential sensitivity between A and B viruses towards
the number of uridine residues overlapping to the RNA duplex
suggests that the polyadenylation signal (Li & Palese, 1994 ;
Luo et al., 1991) provides further specificity. When tested with
influenza A virus, the optimal number of uridine residues was
5–7, consistent with a previous report (Li & Palese, 1994).
However, 7 uridine residues critically reduced the template
activity in influenza B virus. The results agree with RNA
sequence data showing that the number of uridine residues in
influenza B virus RNA is usually 5 or 6, and 7 uridine residues
are not observed adjacent to the RNA duplex. We believe that,
GHJ
Y.-S. Lee and B. L. Seong
although influenza A and B viruses share a common mechanism
in polyadenylation of mRNA, subtle differences in the
polyadenylation signal, both in the length of the RNA duplex
and the U stretch, may play a role in divergence of the two
viruses.
The base pairing in the RNA duplex is required for
initiation of RNA synthesis as proposed in the RNA-fork
model (Fodor et al., 1995 ; Kim et al., 1997). However, to
minimize abortive initiation, the duplex should melt to allow
passage of the RNA polymerase. Likewise, although the same
base pairing is also required for polyadenylation of mRNA
(Luo et al., 1991), melting of the duplex is needed for cRNA
synthesis. These melting processes may depend, in part, on the
processivity of the polymerase. If so, since the duplex region of
the B virus RNA is longer and includes a higher number of base
pairs, as compared to that of A virus RNA, the melting process
may require greater processivity of the influenza virus
polymerase. The rate of evolutionary change is reported to be
higher in A virus than in B virus (Smith & Palese, 1989 ;
Yamashita et al., 1988). It merits further investigation whether
greater processivity of influenza B virus polymerase as
proposed in this study would be related to a lower rate of
nucleotide misincorporation and a lower rate of influenza B
virus evolution.
In summary, the sequence, the number of base pairs and the
number of residues constituting the U stretch in the noncoding
region of the RNA template all affect the influenza A and B
virus transcriptases differentially. It is possible that these
characteristics, in concert, endow differences in specificity
between the two types of influenza virus. While both types of
virus share common structural features, the observed differ-
ential sensitivities towards these cis-acting signals in tran-
scription of the RNA genome, along with unique coding
strategies adopted by the B virus (Horvath et al., 1990 ; Lamb
& Horvath, 1991 ; Shaw et al., 1992), would contribute to the
divergence of these two related viruses (Smith & Palese, 1989 ;
Yamashita et al., 1988). It would be worth testing whether the
same changes in the specificity signals could be introduced into
the endogenous RNA genome and be rescued as transfectant
influenza B virus (Barclay & Palese, 1995). In influenza A virus,
modifications in the noncoding base pairs could be rescued as
infectious viruses, with dramatic changes in virus-specific RNA
synthesis (Bergman & Muster, 1996) or attenuation of
virulence (Muster et al., 1991). A good reverse genetics system
for influenza B virus would be required to dissect the effects of
the specificity signals on virus replication and their potential
effect on attenuation. Judicious choice and combination of
attenuating mutations would result in genetically engineered
influenza B virus suitable for vaccination.
References
Akoto, E. A., Siasubramanian, N. & Nayak, D. P. (1987). Primary
structure of the polymerase acidic (PA) gene of an influenza B virus
(B}SING}222}79). Virology 159, 147–153.
GIA
Almond, J. W., Haymerle, H. A., Felsenreich, V. D. & Reeve, P. (1979).
The structural and infected cell polypeptides of influenza B virus. Journal
of General Virology 45, 611–621.
Barclay, W. S. & Palese, P. (1995). Influenza B viruses with site-specific
mutations introduced into the HA gene. Journal of Virology 69,
1275–1279.
Bergman, M. & Muster, T. (1996). Mutations in the noncoding
sequences of the influenza A virus segments affect viral vRNA formation.
Virus Research 44, 23–31.
DeBorde, D. C., Donabedian, A. M., Herlocher, M. L., Naeve, C. W. &
Maassab, H. F. (1988). Sequence comparison of wild-type and cold-
adapted B}Ann Arbor}1}66 influenza virus genes. Virology 163,
429–443.
Desselberger, U., Racaniello, V. R., Zazra, J. J. & Palese, P. (1980).
The 3« and 5« terminal sequences of influenza virus RNA segments are
highly conserved and show partial inverted complementarity. Gene 8,
315–328.
Enami, M. & Palese, P. (1991). High-efficiency formation of influenza
virus transfectants. Journal of Virology 65, 2711–2713.
Flick, R., Neumann, G., Hoffmann, E., Neumeier, E. & Hobom, G.
(1996). Promoter elements in the influenza vRNA terminal structure.
RNA 2, 1046–1057.
Fodor, E., Pritlove, D. C. & Brownlee, G. G. (1995). Characterization of
the RNA-fork model of virion RNA in the initiation of transcription in
influenza A virus. Journal of Virology 69, 4012–4019.
Fodor, E., Pritlove, D. C., Gould, K. G. & Brownlee, G. G. (1996). The
‘ RNA-fork ’ model for the initiation of influenza transcription. In Options
for the Control of Influenza III, pp. 381–388. Edited by L. E. Brown, A. W.
Hampson & R. G. Webster. Amsterdam : Elsevier.
Gorman, M., Moffat, L. F. & Howard, B. H. (1982). Recombinant
genomes which express chloramphenicol acetyltransferase in mammalian
cells. Molecular and Cellular Biology 2, 1044–1051.
Hagen, M., Chung, T. D. Y., Butcher, J. A. & Krystal, M. (1994).
Recombinant influenza virus polymerase : requirement of both 5« and 3«
viral ends for endonuclease activity. Journal of Virology 68, 1509–1515.
Hall, C. B., Douglas, R. G., Geiman, J. M., Jr & Meagher, M. P. (1979).
Viral shedding patterns of children with influenza B infection. Journal of
Infectious Diseases 140, 610–613.
Horvath, C. M., Williams, M. A. & Lamb, R. A. (1990). Eukaryotic
coupled translation of tandem cistron : identification of the influenza B
virus BM2 polypeptide. EMBO Journal 9, 2639–2647.
Kemdirim, S., Palefsky, J. & Briedis, D. J. (1986). Influenza B virus PB1
protein : nucleotide sequence of the genome RNA segment predicts a
high degree of structural homology with the corresponding influenza A
virus polymerase protein. Virology 152, 126–135.
Kim, H. W., Brandt, C. D., Arrobio, J. O., Murphy, B., Chanock, R. M. &
Parrot, R. H. (1979). Influenza A and B virus infection in infants and
young children during the years 1957–1976. American Journal of
Epidemiology 109, 464–476.
Kim, H.-J., Fodor, E., Brownlee, G. G. & Seong, B. L. (1997).
Mutational analysis of the RNA-fork model of the influenza A virus
vRNA promoter in vivo. Journal of General Virology 78, 353–357.
Kingsbury, D. W. (1990). Orthomyxoviridae and their replication. In
Fields Virology, 2nd edn, pp. 1075–1091. Edited by B. W. Fields & D. M.
Knipe. New York : Raven Press.
Lamb, R. A. & Choppin, P. W. (1983). The gene structure and replication
of influenza virus. Annual Review of Biochemistry 52, 467–506.
Lamb, R. A. & Horvath, C. M. (1991). Diversity of coding strategies in
influenza viruses. Trends in Genetics 7, 261–266.

Lee, Y. S. & Seong, B. L. (1996). Mutational analysis of influenza B
virus RNA transcription in vitro. Journal of Virology 70, 1232–1236.
Li, X. & Palese, P. (1994). Characterization of the polyadenylation
signal of influenza virus RNA. Journal of Virology 68, 1245–1249.
Luo, G., Luytjes, W., Enami, M. & Palese, P. (1991). The poly-
adenylation signal of influenza virus RNA involves a stretch of uridine
followed by the RNA duplex of the panhandle structure. Journal of
Virology 65, 2861–2867.
Luytjes, W., Krystal, M., Enami, M., Parvin, J. D. & Palese, P. (1989).
Amplification, expression, and packaging of a foreign gene by influenza
virus. Cell 59, 1107–1113.
Martı! n, J., Albo, C., Ortı! n, J., Melero, J. A. & Portela, A. (1992). In vitro
reconstitution of active influenza virus ribonucleoprotein complexes
using viral proteins purified from infected cells. Journal of General Virology
73, 1855–1859.
Mena, I., de la Luna, S., Albo, C., Martı! n, J., Nieto, A., Ortı! n, J. &
Portela, A. (1994). Synthesis of biologically active influenza virus core
proteins using a vaccinia virus–T7 RNA polymerase expression system.
Journal of General Virology 75, 2109–2114.
Murphy, B. R. & Webster, R. G. (1990). Orthomyxoviruses. In Fields
Virology, 2nd edn, pp. 1091–1152. Edited by B. N. Fields & D. M. Knipe.
New York : Raven Press.
Muster, T., Subbarao, E. K., Enami, M., Murphy, B. P. & Palese, P.
(1991). An influenza A virus containing influenza B virus 5« and 3«
noncoding regions of the neuraminidase gene is attenuated in mice.
Proceedings of the National Academy of Sciences, USA 88, 5177–5181.
Neumann, G. & Hobom, G. (1995). Mutational analysis of influenza
virus promoter elements in vivo. Journal of General Virology 76,
1709–1717.
Neumann, G., Zobel, A. & Hobom, G. (1994). RNA polymerase I-
mediated expression of influenza viral RNA molecules. Virology 202,
477–479.
Nolan, T. F., Goodman, R. A., Hinman, A. R., Noble, G. R., Kendal, A.
P. & Thacker, S. B. (1980). Morbidity and mortality associated with
influenza B in the United States, 1979–1980. A report from the Centers
for Disease Control. Journal of Infectious Diseases 142, 360.
Piccone, M. E., Fernandez-Sesma, A. & Palese, P. (1993). Mutational
analysis of the influenza virus vRNA promoter. Virus Research 28,
99–112.
Robertson, J. S. (1979). 5« and 3« terminal nucleotide sequences of the
RNA genome segments of influenza virus. Nucleic Acids Research 6,
3745–3757.
Specificity between influenza A and B viruses
Saks, M. E., Sampson, J. R. & Abelson, J. N. (1994). The transfer RNA
identify problem : a search for rules. Science 263, 191–197.
Seong, B. L. & Brownlee, G. G. (1992 a). A new method for
reconstituting influenza polymerase and RNA in vitro : a study of the
promoter elements for cRNA and vRNA synthesis in vitro and viral
rescue in vivo. Virology 186, 247–260.
Seong, B. L. & Brownlee, G. G. (1992 b). Nucleotides 9 to 11 of the
influenza A virion RNA promoter are crucial for activity in vitro. Journal
of General Virology 73, 3115–3124.
Shaw, M. W., Choppin, P. W. & Lamb, R. A. (1983). A previously
unrecognized influenza B virus glycoprotein from a bicistronic mRNA
that also encodes the viral neuraminidase. Proceedings of the National
Academy of Sciences, USA 80, 4879–4883.
Shaw, M. W., Arden, N. H. & Maassab, H. (1992). New aspects of
influenza viruses. Clinical Microbiology Reviews 5, 74–92.
Smith, F. I. & Palese, P. (1989). Variation in influenza virus genes :
epidemiological, pathogenic, and evolutionary consequences. In The
Influenza Viruses, pp. 319–359. Edited by R. M. Krug. New York : Plenum
Press.
Stoeckle, M. Y., Shaw, M. W. & Choppin, P. W. (1987). Segment-
specific and common nucleotide sequences in the noncoding regions of
influenza B virus genome RNAs. Proceedings of the National Academy of
Sciences, USA 84, 2703–2707.
Tiley, L. S., Hagen, M., Matthews, J. T. & Krystal, M. (1994). Sequence-
specific binding of the influenza virus RNA polymerase to sequences
located at the 5« ends of the viral RNAs. Journal of Virology 68,
5108–5116.
Yamanaka, K., Ogasawara, N., Yoshikawa, H., Ishihama, A. & Nagata,
K. (1991). In vivo analysis of the promoter structure of the influenza
virus RNA genome using a transfection system with an engineered RNA.
Proceedings of the National Academy of Sciences, USA 88, 5369–5373.
Yamashita, M., Krystal, M., Fitch, W. M. & Palese, P. (1988). Influenza
B virus evolution : co-circulating lineages and comparison of evolutionary
patterns with those of influenza A and C viruses. Virology 163, 112–123.
Zhang, H. & Air, G. M. (1994). Expression of functional influenza A
polymerase proteins and template from cloned cDNAs in recombinant
vaccinia virus infected cells. Biochemical and Biophysical Research Com-
munications 200, 95–101.
Received 15 September 1997 ; Accepted 28 November 1997
GIB